Topic 4 DQ 1

Contents lists available at ScienceDirect

Preventive Medicine

journal homepage: www.elsevier.com/locate/ypmed

Review Article

Skin cancer interventions across the cancer control continuum: A review of experimental evidence (1/1/2000–6/30/2015) and future research directions

Alan C. Gellera,⁎, Barbra A. Dickermana, Jennifer M. Taberb, Laura A. Dwyerb, Anne M. Hartmanb, Frank M. Pernab

a Harvard TH Chan School of Public Health, United States b National Cancer Institute, United States

A R T I C L E I N F O

Keywords: Skin cancer Intervention Cancer control intervention Behavior Detection Prevention Review Dissemination

A B S T R A C T

While the general efficacy of skin cancer interventions have been reviewed, employing the cancer control continuum would be useful to identify research gaps at specific cancer control points. We characterized the intervention evidence base for specific behavioral targets (e.g., tanning, sun protection, screening) and clinically related targets (e.g., sunburn, skin exams, cancers) at each point in the cancer control continuum. The review included articles published from 1/1/2000–6/30/15 that had an experimental design and targeted behavioral intervention in skin cancer (e.g., specific behaviors or clinically related targets). The search yielded 86 articles, including seven dissemination studies. Of the 79 non-dissemination studies, 57 exclusively targeted primary prevention behaviors, five exclusively targeted screening, 10 targeted both detection and prevention, and eight addressed cancer survivorship. Among prevention studies (n=67), 29 (43%) targeted children and 38 (57%) targeted adults. Of the 15 screening studies, nine targeted high-risk groups (e.g., men aged ≥50 years) and six targeted the general population. Although research has focused on skin cancer prevention, empirically validated interventions are still needed for youth engaged in indoor tanning and for behavioral interventions to pursue change in clinically relevant targets. Research must also address detection among those at highest risk for skin cancer, amelioration of emotional distress attendant to diagnosis and treatment, and survivorship concerns. We discuss essential qualities and opportunities for intervention development and translational research to inform the field.

1. Introduction

The incidence of the most commonly fatal form of skin cancer, cu- taneous melanoma, is increasing faster than any other preventable cancer (Surveillance, Epidemiology, and End Results (SEER) Program, 2015). An estimated 76,000 Americans will be diagnosed with mela- noma in 2015 (American Cancer Society). Adults of all ages are at risk. For men and women ages 20–29, melanoma is one of the most common cancers. Rates have risen precipitously among middle-aged and older men and women. The mortality rate of melanoma has recently stabi- lized, a reflection of decreasing rates of mortality in persons younger than age 60 and sharply increasing rates among people ages 60 and above (Surveillance, Epidemiology, and End Results (SEER) Program, 2015). From 2007 to 2011, nearly 5 million adults were treated for skin cancer annually, with average treatment costs of $8.1 billion (Guy Jr

et al., 2015). The Surgeon General's Call to Action to Prevent Skin Cancer (SG-CTA)

and other publications have identified research gaps for skin cancer prevention intervention (U.S. Department of Health and Human Services, 2014). Improving behavioral intervention and translational research requires identifying research gaps and intervention targets across the cancer control continuum (Lazovich et al., 2012). Con- ceptualizing skin cancer intervention across the cancer control con- tinuum (SCI-3C) (Miller et al., 2009) can characterize the existence, or lack thereof, of intervention efficacy for specific targets across socio- ecological levels of influence (i.e., individual, family, community, and environment/policy level) along the five phases of the cancer control continuum (Fig. 1) (prevention, detection [screening], diagnosis and pre-treatment, treatment, and survivorship). Such models with distinct points in the cancer control continuum exist for physical activity and

https://doi.org/10.1016/j.ypmed.2018.01.018 Received 17 July 2017; Received in revised form 11 January 2018; Accepted 29 January 2018

⁎ Corresponding author at: Department of Social and Behavioral Sciences, Harvard TH Chan School of Public Health, Kresge Building, Room 718, 677 Huntington Avenue, Boston, MA 02115, United States.

E-mail address: [email protected] (A.C. Geller).

Preventive Medicine 111 (2018) 442–450

Available online 06 February 2018 0091-7435/ © 2018 Elsevier Inc. All rights reserved.

T

informatics interventions (Stange et al., 2012; Rowland, 2008; Courneya and Friedenreich, 2007) but none currently exists for skin cancer intervention research. Targets may include specific behaviors (e.g., sun protective practices, skin exams) or clinical (e.g., sunburn), observable (e.g. skin darkening), or patient reported outcomes (e.g., quality of life or distress), but they must reflect the intended result of behavior intervention.

Recently, the Community Preventive Services Taskforce evaluated the effectiveness of seven multicomponent community-wide interven- tions which showed a median increase in sunscreen use of 10.8 (in- terquartile interval = 7.3, 23.2) percentage points, a small decrease in ultraviolet radiation exposure, a decrease in indoor tanning device use of 4.0 (95% CI = 2.5, 5.5) percentage points, and mixed results for other protective behaviors (Sandhu et al., 2016). Given other meta- analyses with much smaller sample sizes that often focus on one com- ponent of the continuum (i.e., prevention), we sought to conduct an extensive, descriptive review of the extant literature with a special emphasis detailing the many components of the continuum. The study was intended as a mapping review in which the statistical significance of individual studies was of less importance than capturing the in- dividual components of given interventions (Grant and Booth, 2009).

The objective of the current review was to describe the state of skin cancer intervention research conducted between 1/1/2000 and 6/30/ 2015 in order to identify research gaps and pressing questions related to specific targets along the cancer control continuum. At each point along the continuum, we group outcomes by research setting and provide information on study features. This characterization of research may communicate research needs to the behavioral research and derma- tology investigative community and foster integration, coordination,

and collaboration among scientists operating at different SCI levels of influence and scientific domains.

2. Methods

2.1. Search strategy and coding procedure

An electronic literature search was performed to identify studies published between 2000 and the first 6 months of 2015. The search strategy was developed in Medline and adapted for PubMed, CINAHL, Embase, and Cochrane Database of Systematic Reviews. We used a combination of MESH terms and keywords listed in the abstracts, in- cluding at least one of the terms: melanoma, skin cancer, skin neoplasm, UV, children, adults, detection, counseling, screening, self-screening, self- examination, dermoscopy, dermatoscopy, early detection, treatment, policy, dissemination, prevention, protection, sun protection, sunscreen, indoor tanning/salons, education, intervention, experimental, trials, randomized. Results were imported into EndNote, where duplicate removal was performed. After initial searches that did not require the words skin cancer to be included, subsequent combination searches, for example, counseling and sun protection, required a skin cancer related term. References used for earlier review articles were examined and a manual search of the reference lists of the retrieved articles was also performed to identify any additional studies.

2.2. Identification of eligible studies

Studies were included in this review if they met the following cri- teria (1) study design: randomized controlled trial or quasi-

Fig. 1. Skin cancer intervention across the cancer control continuum (SCI-3C) conceptual model. The SCI-3C logic model for behavioral intervention research aligns points in the cancer- control and medical time-point continuums and attempts to identify relevant interventional target behaviors (e.g., sun protection). It also links improvements in behavioral outcomes to change in proximal clinically related targets (e.g., eliminate or reduce sunburn) that are reliably associated with skin cancer. For example, behavioral interventions may lead to changes in proximal behaviors that, in turn, affect clinically related targets that are thought to promote distal effects on disease (skin cancer) outcomes. Behavioral interventions may also lead directly to changes in proximal clinical targets (e.g., shade structure additions that does not change behavior but still reduces UV exposure). Lastly, change in some target behaviors, such as indoor tanning, may lead directly to distal skin cancer outcomes. Solid lines depict causal relationships underpinned by experimental evidence has been replicated in at least two studies, and dashed lines represent hypothesized relationships based on epidemiological data.

A.C. Geller et al. Preventive Medicine 111 (2018) 442–450

443

experimental design; (2) intervention: intervention study promoting skin cancer-relevant behaviors and outcomes; (3) outcomes: target de- tection or prevention behaviors included sun safety practices (e.g., midday sun avoidance, shade seeking, protective clothing use, sunsc- reen use, and intentional tanning avoidance [indoor and sunbathing]); (4) published in peer-reviewed journals, and (5) written in the English language. Studies were excluded if they (1) did not use RCT or quasi- experimental design; (2) contained only intermediate outcomes of a behavior (e.g. knowledge or intentions) rather than performance of a behavior, (3) focused on sunscreen efficacy outside of an intervention setting (i.e. validity studies of sunscreen's sun protection factor); (4) reported on results from an existing published trial (or intervention) but used a longer follow up; (5) written in any language other than English, or (6) published prior to 2000 or after 06/30/2015.

One rater working with a research librarian independently reviewed the titles and abstracts of all identified citations in Endnote to match our eligibility criteria.

2.3. Evaluation of eligible studies

For screening studies, we differentiated results for high versus average skin cancer risk groups and arranged studies by population scope. Target behaviors included performing a skin self-exam (SSE) or obtaining a clinical skin exam (CSE). We extracted data on information source (subject versus medical records/clinician report), exam com- prehensiveness (whole, partial, not reported), and inclusion of clin- ician-reported outcomes, skin surgeries, and cost.

Behavioral composite scores were also captured. Clinically-related targets involved reported behavior performance (self/guardian/re- search staff-reported) or an objective outcome measure (nevi, sunburns, skin darkening).

2.4. Data presentation and synthesis

We applied an a priori framework to organize studies: (1) along the cancer control continuum (Fig. 1) based on the proximal and distal targets of the interventions, and (2) across socio-ecological levels of influence (e.g., individual-to-policy based on study population and setting). We distinguished between incidental versus intentional sun exposure and subjective versus objective measures. Dissemination stu- dies were separately characterized. One of us (BD) coded the studies according to the a priori framework and another of us (AG) reviewed a number of these studies using the same criterion.

We report only descriptive statistics due to the large number of possible comparisons and exploratory nature of the review.

3. Results

Using the eligibility criteria noted above, the electronic search identified 518 citations, and after duplicate removal, 495 unique cita- tions were screened for eligibility. 350 were eliminated after title and abstract screening and 75 were eliminated after the full text screening. A total of 70 articles were eligible for this review. The reference lists of retrieved articles and the SG-CTA were manually searched yielding 16 additional articles for a total of 86 studies included in the review meeting eligibility criteria.

Of the 86 articles meeting inclusion criteria, seven were dis- semination studies, which often involved multiple sites as opposed to the single site studies that comprised a sizable share of the remaining 79 studies. Of the remaining 79 articles, 57 reported on interventions that exclusively targeted primary prevention behaviors, 5 exclusively tar- geted screening (e.g., skin examinations), and 10 targeted prevention and screening behaviors. Seven studies focused on survivorship and met the NCI-defined criteria for survivorship research: (1) studies inter- vening on skin-cancer related behaviors among any cancer survivor group, or (2) studies involving family members of cancer survivors and

Fig. 2. Frequency of intervention studies (1/1/2000–6/30/2015) reporting skin cancer prevention behaviors among children and intervention setting.

A.C. Geller et al. Preventive Medicine 111 (2018) 442–450

444

intervening on a family member's skin cancer-related behavior.

3.1. Prevention childhood

Of the 67 studies targeting skin cancer prevention, 29 targeted child behaviors (43%) and 38 targeted adult behaviors (57%) (see Figs. 2 & 3). Among the 29 childhood prevention studies, the distribution of population scope was: school setting (52%), community/recreation setting (21%), parents of children (17%), and medical/special popula- tions (10%). Intervention duration ranged from < 1 day to 4 years, with the majority of studies (45%) in the 0–3 month range (mean 13.5 months; median 4.5 months). Duration of the longest follow-up ranged from 5 weeks to 6 years, with most (31%) in the 0–3 month range (mean 17.8 months; median 12 months). The number of partici- pants ranged from 80 to 8721 (mean 1688; median 1172).

Measured sun protective behaviors included: sunscreen use (n = 23), protective clothing use (n = 17), hat use (n = 17), shade use (n = 14), sunglasses use (n = 9), midday sun avoidance (n = 8), minimizing incidental sun exposure (n = 2), sunbathing avoidance (n = 1), and indoor tanning avoidance (n = 0). The number of studies that obtained significant behavioral outcomes was as follows: sunscreen use (n = 10), shade use (n = 7), hat use (n = 5), protective clothing use (n = 4), sunglasses use (n = 3), and midday sun avoidance (n = 2). Of the 14 of 29 studies (48%) that included a composite score of protective behaviors, 13 achieved a significant change in this index and 1 did not. To assess behavioral change, 15 studies measured self-reported beha- vior, 11 measured guardian-reported behavior, and 5 measured re- search staff-reported behavior (total > 29 due to use of more than one type of report in some studies).

Fifty-nine percent (17 of 29) included at least one measure of an objective clinically-related target (sunburn frequency [n = 11], change in nevi [n = 7], and/or skin darkening [n = 5]). Significant outcomes were attained for sunburn frequency (n = 4) and skin darkening

(n = 1).

3.2. Prevention adulthood

Among the 38 adulthood prevention studies, the distribution of population scope was: young adults/intentional tanners (28%), general population (18%), medical setting (13%), recreational setting (13%), special populations (8%), occupational setting (10%), and mis- cellaneous (8%). Intervention duration ranged from < 1 day to 3 years, with the majority (76%) in the 0–3 month range (mean 3.5 months; median 19 days). Duration of the longest follow-up ranged from 2 weeks to 3 years, again with most (50%) in the 0–3 month range (mean 7.6 months; median 3.5 months). The number of participants ranged from 50 to 8295 (mean 1065; median 197).

Measured protective behaviors included: sunscreen use (n = 21), hat use (n = 13), sunbathing avoidance (n = 12), indoor tanning avoidance (n = 8), protective clothing use (n = 9), midday sun avoid- ance (n = 7), shade use (n = 6), sunglasses use (n = 7), and minimizing incidental sun exposure (n = 6). The number of studies that obtained significant behavioral outcomes were: sunscreen use (n = 13), hat use (n = 6), sunbathing avoidance (n = 4), sunglasses use (n = 4), indoor tanning avoidance (n = 5), midday sun avoidance (n = 2), and mini- mizing incidental sun exposure (n = 2). Of the 18 of 38 studies (47%) that included a composite score of protective behaviors, 13 achieved a significant change and 5 did not. All 38 studies used self-reported be- havior to assess behavioral change.

In addition to reported behavioral change, 14 studies (37%) in- cluded at least one measure of an objective outcome (sunburn fre- quency [n = 12] and/or skin darkening [n = 3]; note: total > 13 due to the fact that one study included measures of both). There were 4 in- stances of a significant impact on sunburn frequency and 2 instances of a significant impact on skin darkening (note: these findings spanned 5 separate studies as 1 study reported on both of these objective

Fig. 3. Frequency of intervention studies (1/1/2000–6/30/2015) reporting skin cancer prevention behaviors among adults and intervention setting.

A.C. Geller et al. Preventive Medicine 111 (2018) 442–450

445

measures).

3.3. Prevention-related dissemination trials

Three dissemination trials (with seven publications) have been conducted across US pools (Glanz et al., 2015), ski areas (Walkosz et al., 2014; Walkosz et al., 2015), and zoos (Lewis et al., 2005). Interventions have generally included education of participants (employees), built environment improvement, and provision of incentives. Two of the three trials tested the use of an enhanced dissemination strategy com- pared with basic dissemination. Go Sun Smart compared enhanced vs. basic strategies, and reported similar effects for sun safety (Walkosz et al., 2015). Pool Cool employed a theory-driven intervention in en- hanced sites that included incentives, reinforcement, feedback, and skill-building for pool staff. Initially, both basic and enhanced condi- tions achieved significant policy and environmental improvements, with non-significant differences between conditions. In year two and at last follow-up, the enhanced condition improved sun safety practices and environments more than the basic group and had greater policy implementation and maintenance. Finally, the dissemination studies uniformly conducted process and outcome evaluation at multiple sites.

3.4. Early detection

Among the 15 early detection studies, 9 targeted groups known to be above population risk for melanoma (age 50 and older, NMSC pa- tients [i.e. transplant], and those identified as high risk per brief skin cancer risk assessment tool [BRAT] score) while 6 targeted members of the general population (see Fig. 4). The distribution of population scope was: medical patients who do not have a melanoma diagnosis (47%), general population (13%), outdoor workers (13%), beachgoers (7%), primary care office patients (7%), those ages 50 or older (7%), high risk per BRAT score (7%). Intervention duration ranged from < 1 day to 1.75 years, with the majority (n = 10) in the 0–3 month range (mean 3.6 months; median < 1 day). Follow-up duration ranged from 1 month to 1.75 years, with the majority (n = 5) in the 0–3 month range (mean 7.2 months; median 5 months). The number of participants ranged from 75 to 1356 (mean 494; median 495).

Measures of skin examination were most commonly based on par- ticipant self-reported skin self-exam (SSE) (n = 16), followed by parti- cipant self-reported clinical skin exam (CSE) (n = 3). Least common was CSE performance obtained from clinician report or medical records (n = 1). We further extracted data on the comprehensiveness of the skin examination (whole-body skin check, partial-body skin check, not reported). Among the 16 measures of self-reported SSE performance, 10 did not specify exam comprehensiveness, 3 involved whole-body skin checks, and 3 involved partial-body checks.

Two studies included at least one measure of an objective outcome (clinician-reported CSE outcomes [n = 1], skin surgeries [n = 2], and/ or cost of CSE [n = 1]). There was one instance of a significant inter- vention impact on clinician-reported CSE outcomes and one instance of a significant impact on skin surgeries, which were both reported in the same study targeting men age 50 and older (Janda et al., 2014).

3.5. Survivorship

Eight articles pertained to survivorship. Only two of these studies targeted melanoma survivors' behavior specifically, and both inter- ventions were aimed at promoting SSEs (Janda et al., 2014; Robinson et al., 2007). Four of the papers reported on interventions that targeted prevention and screening behaviors of melanoma survivors' relatives: 3 targeted family members broadly (Robinson et al., 2010; Manne et al., 2010; Bowen et al., 2015), 1 targeted siblings (Geller et al., 2006) and 1 targeted prevention behaviors only among melanoma survivors' chil- dren (Gritz et al., 2013). Only 1 study focused on prevention behaviors among adolescent survivors of childhood cancer (not restricted to melanoma diagnosis) (Mays et al., 2011). All 8 studies showed modest effects, including two for sun safety in adolescents and the remainder for SSE among adults.

4. Discussion

This extensive review of the past 16 years of experimental beha- vioral research in skin cancer along the cancer control continuum highlights the growth and diversity of interventions in multiple settings, but also reveals glaring gaps along both the cancer control continuum and the translational research continuum. Overall, the large majority of studies were at the prevention point along the cancer control con- tinuum with far less research aimed at the detection (18%), diagnosis/ pre-treatment (4%), treatment (0%), and survivorship (8%) points on the continuum.

4.1. Prevention

Of the 67 studies targeting skin cancer prevention, 57% were in various adult settings, and were generally of brief duration (e.g., 74% 0–3 months). Pediatric studies (43%) were also of short duration, and occurred across fewer settings, with schools being the most common intervention site. While many studies reported statistically significant results suggesting favorable intervention effects, a closer inspection revealed important gaps. For example, only one pediatric study ad- dressed indoor tanning behavior (Mays et al., 2011). While some chil- dren (i.e., under age 21) are captured in studies of college students, an estimated 25–30% of teen girls report indoor tanning (Aarestrup et al.,

Fig. 4. Frequency of intervention studies (1/1/2000–6/30/2015) reporting skin cancer screening by skin exam type and intervention setting.

A.C. Geller et al. Preventive Medicine 111 (2018) 442–450

446

2014) and regulation banning indoor tanning among children < 18- years is still pending. Further research should determine the potential moderating effects of state and local policy environments on interven- tion efficacy, particularly for indoor and outdoor tanning among teens and young adults.

This review and SCI-3C logic model highlights the need for research to demonstrate effects beyond achieving significant change in target behaviors (e.g., sun protection) toward improvements in clinically-re- lated proximal targets (e.g., reduction sunburn/tanning). For example, only 58% and 33% of childhood and adulthood studies reviewed, re- spectively, included a clinically-related target, and of those, a third or fewer of these interventions demonstrated a significant effect on the proximal target. Unlike other cancer-related health behaviors, an evi- dence-based minimally effective dose of sun protection behavior re- mains elusive. While limited by its cross-sectional design, the most re- cent national study of adults reported that use of sun protective behavior was modestly related to sunburn reduction among White non- Hispanics (p = .056) (Holman et al., 2014). However, with few ex- ceptions (Green et al., 2011; Gallagher et al., 2000; van der Pols et al., 2006) reported sun-protective behaviors have not been reliably related to clinically-relevant outcomes, such as sunburn and melanoma re- duction. Future intervention research is needed to demonstrate potency to favorably affect proximal clinically-related targets (e.g., reduced or zero sunburn). Studies to improve methodology and measurement sensitivity of behavioral and clinically-related targets are also im- portant. This proposed research direction does not negate the initial success of intervention approaches to change target behaviors, but ra- ther emphasizes that increasing intervention potency and including proximal intervention targets may be needed to improve distal skin cancer related outcomes.

Despite involvement in multiple settings, this review underscores that there has been little headway in conducting multi-level interven- tions. In contrast, in a summary of the first 20 years (1980–2000) of activities in the Australian SunSmart program, Montague et al. (2001) described successful coordinated outreach with a strong state and na- tional evaluation, activity and interventions across many levels (in- dividuals, families, organizations, and the community at large), in many settings (education, workplace, sport and recreation, fashion in- dustry, community health, and local government), and with many target groups (children, young adults, parents and teachers, older adults, outdoor workers, nurses and general practitioners). However, as compared with the efforts of the comprehensive Australian SunSmart program and US dissemination studies, interventions reported herein often lack coordination, exist primarily at single-sites, rarely connect level, setting, and target group, and do not always have a comprehen- sive evaluation plan. Interventions may include collaborative efforts between schools, workplaces and recreational settings, with aims to (1) optimize physical and policy environments for shade, protective clothing, and sunscreen, (2) involve public reminders for their use in multiple settings, and (3) leverage individually tailored, real-time communication in this work.

4.2. Detection

Compared with primary prevention interventions, significantly fewer randomized studies (n = 15) centered on screening and early detection. While two-thirds of studies targeted populations at higher risk of melanoma, only one study (Janda et al., 2014) focused on middle-aged and older white men, the group at greatest risk of fatal melanoma (Surveillance, Epidemiology, and End Results (SEER) Program, 2015). Approximately two-thirds reported significant results for completion of any type of skin exam, but relatively few included objective outcomes or costs. Set against these trials in smaller settings, population-based and large workplace educational programs and screenings have taken place in Germany and at the Lawrence Livermore National Laboratory, respectively, targeting clinical outcomes and

screening behavior (Katalinic et al., 2012; Schneider et al., 2008). While the Schleswig-Holstein (Germany) evaluation reported initially promising clinical outcomes, longer term follow-up found mortality rates to be equal to those at baseline (Boniol et al., 2015). Further be- havioral intervention research must causally relate behavior change to improvement in clinically-related targets while minimizing potential risks. The NCI recommends annual skin cancer screening for specific groups (Late Effects of Treatment for Childhood Cancer (PDQ®), 2012) and research to improve intervention potency to hit behavioral and clinical targets within these groups is needed. Further, intervention studies involving groups for whom skin screening is recommended could be used to introduce and evaluate new methodology to improve sensitivity, specificity, and collection of harm data that could improve and inform the conduct of future intervention research with high risk groups for whom no recommendation for skin cancer screening cur- rently exists.

4.3. Cancer survivorship

The paucity of behavioral intervention research for melanoma sur- vivors at the UV protective behavior portion of the continuum is striking in comparison to other cancers for which behavioral inter- vention has been deployed to improve treatment compliance, amelio- rate psychological distress and physical symptomatology, and generally improve quality of life (Rowland, 2008; Courneya and Friedenreich, 2007). However, little is known about the psychological distress that may accompany melanoma and NMSC diagnosis or treatment. The advancement and variability of pharmacological treatment of mela- noma suggests that survivorship research will become increasingly important, and disfigurement and recurrence of NMSC may also war- rant behavioral intervention for which there is little evidence.

The following sections build off of the current review to highlight SCI-3C research tasks and conclude with a series of questions for further research.

4.4. Lack of intervention potency, breadth, and sustainability in UV protection

Future research must first explore why UV protection efforts have lacked potency, breadth, and sustainability. First, sun protection mes- saging likely suffers from an inability to home in on a single sun pro- tection message because of the wide variety of ways that the public can protect themselves from the sun. Future communication research must strive to determine the viability of venue-specific targeted messages such as alternatives to natural shading at sun-drenched beaches where little shading is provided. Second, the public's perceptions of the risks and benefits of intermittent or incidental sun exposure should continue to be explored, as has been done with Go Sun Smart studies focused on reducing ski slope guests' exposure to winter and intermittent sun ex- posure. Two recent studies report an unintended sunburn frequency of 33% in Denmark and a rate of 13% for just the prior weekend in New Zealand. In the latter study, those who received unintended sunburn were more likely to have been near water and in unshaded areas, and were outdoors for a longer period of time with less body coverage. The authors noted that as sunburn was unintended for these respondents outdoor sun protective behaviors may be amenable to change (McLeod et al., 2017; Koster et al., 2010).

Third, a greater connection between a ‘schools and pools’ approach must be investigated – at the community or state level, how can a yearlong, sustainable commitment to education around the importance of sun protection, the skills required to practice optimal sun protection, and the resources to do so be fostered? Fourth, despite far greater ex- pansion of sun protection into multiple outdoor venues over the past 15 years, compared with earlier periods, there has been little effort to understand the lack of depth – namely what are the lessons that can be drawn from more disseminable activities at schools, pools, ski slopes,

A.C. Geller et al. Preventive Medicine 111 (2018) 442–450

447

and zoos for interventions that have not gone beyond single venues? Namely, we call and advocate for meetings, forums, and workshops to expand the dissemination and implementation capacity of the majority of skin cancer studies that have often resided in single settings.

4.5. Integrating sun protection within outdoor settings and exercise

There are a number of research tasks and public health approaches to intertwine UV protection efforts within outdoor exercise and re- creation settings including improved understanding of relevant signage, provision of sunscreen dispensers, facilitating relationships with pre- viously unexplored industry such as travel and amusement parks, and exploring incorporation of sun protection messaging into highly used technologically-driven exercise tools such as Fitbit.

4.6. Measurement

Glanz and colleagues argue that further quantitative evaluation could be conducted to evaluate internal consistency, test-retest relia- bility, and concurrent and criterion validity (e.g., by comparing the current items to objective measures such as observation, skin re- flectance, personal dosimetry, skin swabbing, and inspection of moles) (Glanz and Mayer, 2005; Glanz et al., 2008). Agreement on the wording for self-report measures and employing more consistent sun protection indices would be of vital importance. At the level of measurement, more objective measurements including observation of sun protection prac- tices would be of value (e.g. measuring how much sun protection is provided). Experts and advocates must also strive to routinely include questions on sun exposure, sun protection, and tanning bed use in na- tional surveys, including Youth Risk Behavioral Surveillance System and Behavioral Risk Factor Surveillance System. A more complete measure of SSE, including specific body sites examined, would be a meaningful advance. Finally, as noted earlier, assessments, particularly those that are longitudinal in nature, must tie behavior in to achieved health outcomes.

In its most nascent stage, research must include a 360 degree ana- lysis of deficits in survivorship research/practice (Rock et al., 2012) – what are the needs of skin cancer survivors, what are physician screening practices regarding survivors, and how are health care sys- tems aligned to alert providers to screening/counseling for skin cancer patients. In particular, health policy research could be bolstered to determine the proportion of melanoma survivors who receive at least annual full skin examinations (and for how long after their initial di- agnosis), perform thorough SSEs, and practice vigilant sun protection. Equally as important, none of the survivorship papers reviewed herein satisfied the following criteria: (1) studies intervening on behavior re- lated to skin cancer-related treatment adherence or quality of life (QOL) (physical or emotional), or (2) studies intervening on QOL/emotion targets of the family members of cancer survivors.

With the advent of new treatments for metastatic melanoma that for some individuals are now showing efficacy, public health strategies building on clinical advances are required and include: ensuring easy access to treatment for persons of low socioeconomic status, uninsured, and underinsured; increasing awareness of the availability of various treatments among internists and community oncologists; improving communication of the benefits and risks of various treatments by on- cologists as they discuss treatment options, measuring patient percep- tion of prognosis and impact on QOL, and seeking ways to improve treatment adherence.

4.7. Accounting for social class, race, and ethnicity

While African Americans and Hispanics relative to whites dis- proportionately die from melanoma (Byrd et al., 2004; Bellows et al., 2001), incidence and mortality rates are 20–50 times greater for whites (Surveillance, Epidemiology, and End Results (SEER) Program, 2015).

Among whites, there is an important social class differential. Whites of low SES, compared with high SES, are more likely to be diagnosed with late-stage melanoma and die of their disease (Pollitt et al., 2011; Geller et al., 1996; Reyes-Ortiz et al., 2006). One study of 566 cases found that lower SES melanoma patients (99% white) were less aware of mela- noma and less worried about the consequences of the disease and more unlikely to ask to be screened for melanoma in the year prior to diag- nosis (Swetter et al., 2012). Future research must seek to improve risk communication for early detection with a stronger health literacy per- spective as well as to explore the role of social class for reasons why less income may be associated with poorer use of sunscreen and why indoor tanning salons are located in low SES white areas.

We conclude with a series of research questions and call for a forum to discuss these and other questions of interest.

4.8. Research questions

4.8.1. Prevention

1. What level of target behavior is needed to produce a significant change in proximal clinical target associated with distal outcome, or alter disease outcome? How does this differ for those at average risk vs. high risk?

2. How can interventions be optimized by including multilevel ap- proaches necessary to hit clinical-related targets? How can these interventions be tailored for groups at high risk for UV exposure, or those with greater skin cancer risk?

3. Considering that unintended sun exposure often resulting in sun- burn occurs in the context of other behaviors such as outdoor phy- sical activity and alcohol consumption, how may skin cancer in- tervention approaches fit and be effective in a multiple behavior framework?

4.8.2. Screening

1. Current guidelines call for regular skin exams for specific groups, such as organ transplant patients and childhood cancer survivors, but they may be underutilized. What interventions may be suc- cessful at targeting these sub-populations?

2. Currently, insufficient evidence exists for recommending general population skin cancer screening, but other groups at high risk of fatal disease (such as middle-aged and older men) may benefit from screening. What are the best practices to improve cancer outcomes and reduce iatrogenic effects?

4.8.3. Diagnosis and survivorship

1. Emotional distress is attendant to cancer diagnoses and medical treatment, and behavioral interventions have been shown to reduce distress and improve behaviors that may augment treatment but no research has addressed these issues in the context of melanoma. What is the severity and trajectory of emotional distress in newly diagnosed melanoma patients?

2. What is the most appropriate course of intervention? 3. May behavioral intervention support other behaviors to improve

medical outcomes related to acute or post-operative treatment?

5. Conclusion

Our review indicates a number of research gaps along the cancer control continuum, especially intervention research addressing detec- tion in high-risk populations, amelioration of emotional distress atten- dant to diagnosis and treatment, and survivorship concerns. While the majority of research concerned prevention, there remains a need to develop interventions for youth engaged in indoor tanning, and in general, to demonstrate that behavioral change leads to change in

A.C. Geller et al. Preventive Medicine 111 (2018) 442–450

448

clinically related targets. Future behavioral intervention research is also needed to address cross-cutting issues related to measurement sensi- tivity, consideration of race and social class on risk perception and communication, and potential uses of social networks that may inform future behavioral intervention development for skin cancer prevention and control.

Disclaimer

The opinions in this paper are those of the authors and do not re- present the NCI or US Government.

Statement of all funding sources for the work

This work was partially supported by contract HHSN2612012000028I to the Westat Corporation and a Westat sub- contract to Alan Geller.

Financial disclosure

None.

Conflict of interest statement

ACG, BAD, JMT, LAD, AMH, and FMP report that they have no conflict of interest.

Acknowledgement

We wish to acknowledge Tracey Goldner, an NCI Behavioral Research Program communications fellow, for the development of fig- ures used in this report.

Appendix A. Supplementary data

Supplementary data to this article can be found online at https:// doi.org/10.1016/j.ypmed.2018.01.018.

References

Aarestrup, C., Bonnesen, C.T., Thygesen, L.C., Krarup, A.F., Waagstein, A.B., Jensen, P.D., Bentzen, J., 2014. The effect of a school-based intervention on sunbed use in Danish pupils at continuation schools: a cluster-randomized controlled trial. J. Adolesc. Health 54 (2), 214–220.

American Cancer Society http://www.cancer.org/cancer/skincancer-melanoma/ detailedguide/melanoma-skin-cancer-key-statistics, Accessed date: 7 July 2017.

Bellows, C.F., Belafsky, P., Fortgang, I.S., Beech, D.J., 2001. Melanoma in African- Americans: trends in biological behavior and clinical characteristics over two dec- ades. J. Surg. Oncol. 78, 10–16.

Boniol, M., Autier, P., Gandini, S., 2015. Melanoma mortality following skin cancer screening in Germany. BMJ Open 15 (5(9)), e008158.

Bowen, D.J., Burke, W., Hay, J.L., Meischke, H., Harris, J.N., 2015. Effects of web-based intervention on risk reduction behaviors in melanoma survivors. J. Cancer Surviv. 9 (2), 279–286.

Byrd, K.M., Wilson, D.C., Hoyler, S.S., Peck, G.L., 2004. Advanced presentation of mel- anoma in African Americans. J. Am. Acad. Dermatol. 50 (1), 21–24 (discussion 142-3).

Courneya, K.S., Friedenreich, C.M., 2007. Physical activity and cancer control. Semin. Oncol. Nurs. 23, 242–252.

Gallagher, R.P., Rivers, J.K., Lee, T.K., Bajdik, C.D., McLean, D.I., Coldman, A.J., 2000. Broad-spectrum sunscreen use and the development of new nevi in white children: a randomized controlled trial. JAMA 14 (283(22)), 2955–2960.

Geller, A.C., Miller, D.R., Lew, R.A., Clapp, R.W., Wenneker, M.B., Koh, H.K., 1996. Cutaneous melanoma mortality among the socioeconomically disadvantaged in Massachusetts. Am. J. Public Health 86 (4), 538–543.

Geller, A.C., Emmons, K.M., Brooks, D.R., et al., 2006. A randomized trial to improve early detection and prevention practices among siblings of melanoma patients. Cancer 107, 806–814.

Glanz, K., Mayer, J.A., 2005 Aug. Reducing ultraviolet radiation exposure to prevent skin cancer methodology and measurement. Am. J. Prev. Med. 29 (2), 131–142.

Glanz, K., Yaroch, A.L., Dancel, M., Saraiya, M., Crane, L.A., Buller, D.B., Manne, S., O'Riordan, D.L., Heckman, C.J., Hay, J., Robinson, J.K., 2008. Measures of sun ex- posure and sun protection practices for behavioral and epidemiologic research. Arch.

Dermatol. 144 (2), 217–222. Glanz, K., Escoffery, C., Elliott, T., Nehl, E.J., 2015. Randomized trial of two dis-

semination strategies for a skin cancer prevention program in aquatic settings. Am. J. Public Health 105 (7), 1415–1423.

Grant, M.J., Booth, A., 2009. A typology of reviews: an analysis of 14 review types and associated methodologies. Health Inf. Libr. J. 26, 91–108.

Green, A.C., Williams, G.M., Logan, V., Strutton, G.M., 2011. Reduced melanoma after regular sunscreen use: randomized trial follow-up. J. Clin. Oncol. 20 (29(3)), 257–263.

Gritz, E.R., Tripp, M.K., Peterson, S.K., Prokhorov, A.V., Shete, S.S., Urbauer, D.L., Fellman, B.M., Lee, J.E., Gershenwald, J.E., 2013. Randomized controlled trial of a sun protection intervention for children of melanoma survivors. Cancer Epidemiol. Biomark. Prev. 22 (10), 1813–1824.

Guy Jr., G.P., Machlin, S.R., Ekwueme, D.U., Yabroff, K.R., 2015. Prevalence and costs of skin cancer treatment in the U.S., 2002–2006 and 2007–2011. Am. J. Prev. Med. 48 (2), 183–187.

Holman, D.M., Berkowitz, Z., Guy Jr., G.P., Hartman, A.M., Perna, F.M., 2014 Jun. The association between demographic and behavioral characteristics and sunburn among U.S. adults - National Health Interview Survey, 2010. Prev. Med. 63, 6–12.

Janda, M., Youl, P., Neale, R., Aitken, J., Whiteman, D., Gordon, L., Baade, P., 2014. Clinical skin examination outcomes after a video-based behavioral intervention: analysis from a randomized clinical trial. JAMA Dermatol. 150 (4), 372–379.

Katalinic, A., Waldmann, A., Weinstock, M.A., Geller, A.C., Eisemann, N., Greinert, R., Volkmer, B., Breitbart, E., 2012 Nov 1. Does skin cancer screening save lives?: an observational study comparing trends in melanoma mortality in regions with and without screening. Cancer 118 (21), 5395–5402.

Koster, B., Thorgaard, C., Phillip, A., et al., 2010. Prevalence of sunburn and sun-related behavior in the Danish population: a cross-sectional study. Scand. J. Public Health 38, 548–552.

Late Effects of Treatment for Childhood Cancer (PDQ®) [Website]. Bethesda, MD: National Cancer Institute; 2012. Available from: http://www.cancer.gov/ cancertopics/pdq/treatment/lateeffects/HealthProfessional/page2.

Lazovich, D., Choi, K., Vogel, R.I., 2012. Time to get serious about skin cancer prevention. Cancer Epidemiol. Biomark. Prev. 21, 1893–1901.

Lewis, E., Mayer, J.A., Slymen, D., Belch, G., Engelberg, M., Walker, K., Kwon, H., Elder, J., 2005. Disseminating a sun safety program to zoological parks: the effects of tai- loring. Health Psychol. 24 (5), 456–462.

Manne, S., Jacobsen, P.B., Ming, M.E., Winkel, G., Dessureault, S., Lessin, S.R., 2010. Tailored versus generic interventions for skin cancer risk reduction for family members of melanoma patients. Health Psychol. 29 (6), 583–593.

Mays, D., Black, J.D., Mosher, R.B., Shad, A.T., Tercyak, K.P., 2011. Improving short-term sun safety practices among adolescent survivors of childhood cancer: a randomized controlled efficacy trial. J. Cancer Surviv. 5 (3), 247–254.

McLeod, G.F.H., Reeder, A.I., Gray, A.R., McGee, R., 2017. Unintended sunburn: a po- tential target for sun protection messages. J. Skin Cancer 2017, 6902942. http://dx. doi.org/10.1155/2017/6902942. Epub 2017 Apr 3.

Miller, S.M., Bowen, D.J., Croyle, R.T., Rowland, J.H. (Eds.), 2009. Handbook of Cancer Control and Behavioral Science: A Resource for Researchers, Practitioners, and Policymakers. American Psychological Association, Washington, DC. http:// cancercontrol.cancer.gov/OD/continuum.html.

Montague, M., Borland, R., Sinclair, C., 2001. Slip! Slop! Slap! and SunSmart, 1980–2000: skin cancer control and 20 years of population-based campaigning. Health Educ. Behav. 28 (3), 290–305.

Pollitt, R.A., Clarke, C.A., Swetter, S.M., Peng, D.H., Zadnick, J., Cockburn, M., 2011. The expanding melanoma burden in California Hispanics: importance of socioeconomic distribution, histologic subtype, and anatomic location. Cancer 1 (117(1)), 152–161.

van der Pols, J.C., Williams, G.M., Pandeya, N., Logan, V., Green, A.C., 2006. Prolonged prevention of squamous cell carcinoma of the skin by regular sunscreen use. Cancer Epidemiol. Biomark. Prev. 15 (12), 2546–2548.

Reyes-Ortiz, C.A., Goodwin, J.S., Freeman, J.L., Kuo, Y.F., 2006. Socioeconomic status and survival in older patients with melanoma. J. Am. Geriatr. Soc. 54, 1758–1764.

Robinson, J., Turrisi, R., Stapleton, J., 2007. Efficacy of a partner assistance intervention designed to increase skin self-examination performance. Arch. Dermatol. 143, 37–41.

Robinson, J.K., Turrisi, R., Mallett, K., Stapleton, J., Pion, M., 2010. Comparing the ef- ficacy of an in-person intervention with a skin self-examination workbook. Arch. Dermatol. 146 (1), 91–94.

Rock, C.L., Doyle, C., Demark-Wahnefried, W., Meyerhardt, J., Courneya, K.S., Schwartz, A.L., Bandera, E.V., Hamilton, K.K., Grant, B., McCullough, M., Byers, T., Gansler, T., 2012 Jul-Aug. Nutrition and physical activity guidelines for cancer survivors. CA Cancer J. Clin. 62 (4), 243–274 (Erratum in: CA Cancer J Clin. 2013 May;63(3):21).

Rowland, J., 2008. Cancer survivorship: rethinking the cancer control continuum. Semin. Oncol. Nurs. 24, 145–152.

Sandhu, P.K., Elder, R., Patel, M., Saraiya, M., Holman, D.M., Perna, F., Smith, R.A., Buller, D., Sinclair, C., Reeder, A., Makin, J., McNoe, B., Glanz, K., Community Preventive Services Task Force, 2016. Am. J. Prev. Med. 51 (4), 531–539.

Schneider, J.S., Moore 2nd, D.H., Mendelsohn, M.L., 2008. Screening program reduced melanoma mortality at the Lawrence Livermore National Laboratory, 1984 to 1996. J. Am. Acad. Dermatol. 58 (5), 741–749.

Stange, K.C., Breslau, E.S., Dietrich, A.J., Glasgow, R.E., 2012. State-of-the-art and future directions in multilevel interventions across the cancer control continuum. J. Natl. Cancer Inst. Monogr. (44), 20–31.

Surveillance, Epidemiology, and End Results (SEER) Program, 2015. (www.seer.cancer. gov). SEER*Stat Database: Incidence - SEER 9 Regs Research Data, Nov 2014 Sub (1973–2012) < Katrina/Rita Population Adjustment > - Linked to County Attributes - Total U.S., 1969–2013 Counties. National Cancer Institute, DCCPS, Surveillance Research Program, Surveillance Systems Branch released April 2015, based on the

A.C. Geller et al. Preventive Medicine 111 (2018) 442–450

449

November 2014 submission. http://seer.cancer.gov/statfacts/html/melan.html. Swetter, S.M., Pollitt, R.A., Johnson, T.M., Brooks, D.R., Geller, A.C., 2012. Behavioral

determinants of successful early melanoma detection: role of self and physician skin examination. Cancer 118 (15), 3725–3734.

U.S. Department of Health and Human Services, 2014. The Surgeon General's Call to Action to Prevent Skin Cancer. U.S. Dept of Health and Human Services, Office of the Surgeon General, Washington, DC.

Walkosz, B.J., Buller, D.B., Andersen, P.A., Scott, M.D., Dignan, M.B., Cutter, G.R., Liu, X., Maloy, J.A., 2014 Sep. Dissemination of go sun smart in outdoor recreation: effect of program exposure on sun protection of guests at high-altitude ski areas. J. Health Commun. 19 (9), 999–1016.

Walkosz, B.J., Buller, D.B., Andersen, P.A., Scott, M.D., Liu, X., Cutter, G.R., Dignan, M.B., 2015. Translation of a ski school sun safety program to North American ski and snowboard schools. Health Promot. Pract. 16 (4), 560–570.

A.C. Geller et al. Preventive Medicine 111 (2018) 442–450

450